The net radiation absorbed by the continents is partitioned mainly into sensible and latent (evapotranspiration) heat fluxes whose release back into the atmosphere directly influences local air temperature and humidity and thence other climate system variables. In any given locale, soil moisture availability and vegetation state largely determine the fraction of net radiation that is used for evapotranspiration, as well as the photosynthetic and respiration rates. Thus, realistic modelling of land surface-atmosphere interactions is essential to realistic prediction of continental climate and hydrology. In doing this, attention must be paid to the links between vegetation and the terrestrial energy, water and carbon cycles, and how these might change due to eco-physiological responses to elevated CO2 and changes in land use.
The exchanges of energy, momentum, water, heat and carbon between the land surface and the atmosphere must be more realistically and accurately calculated in the next generation of coupled models (Sellers et al., 1997). Fluxes of the first four quantities, traditionally defined as physical climate system variables, are routinely parametrized in Numerical Weather Prediction (NWP) models and climate models as functions of the surface albedo, aerodynamic roughness length and surface "moisture availability" (Betts et al., 1998; Viterbo et al., 1999). These land-surface properties can all be defined as functions of the type and density of the local vegetation, and the depth and physical properties of the soil. The first generation of land-surface parametrizations (LSPs) developed in the 1970s took little account of these relationships, and were replaced by biophysically realistic models, complete with supporting vegetation and soil databases, which led to significant improvements in NWP and climate model performance in the 1980s and early 1990s. However, these second generation models incorporated only empirical descriptions of the evapotranspiration process, by which water is taken up from the soil by plant roots and released into the atmosphere through tiny pores in leaf surfaces called stomata, while CO2 is drawn from the atmosphere into leaf interiors for photosynthesis through these same stomata. Research has shown that living plants appear to actively control stomatal widths (conductance) in response to changes in water vapour and CO2 concentration to optimise the ratio of water vapour losses to CO2 uptake, and simple, robust models of the photosynthesis-conductance system in plant leaves have been constructed based on this idea, see Figure 7.5 (Farquhar et al., 1980; Collatz et al., 1991; Sellers et al., 1992a). These models have been parametrized and verified at the leaf level, and can also be scaled up to describe vegetation canopy processes at regional scales using satellite data. These third generation LSPs, published in the late 1990s, thus combine consistent descriptions of the physical climate system transfer processes for energy, momentum, water and heat, with the biophysics of photosynthesis (Bonan, 1995; Sellers et al., 1996c; Dickinson et al., 1998). Why is this important?
Photosynthesis and respiration are climatically sensitive and exhibit interannual variations following climate variations (Dai and Fung, 1993; Francey et al., 1995; Goulden et al., 1996; Myneni et al., 1997; Randerson et al., 1997; Xiao et al., 1998; Randerson et al., 1999; Tian et al., 1999). Furthermore, there appears to have been a net enhancement of terrestrial photosynthesis over respiration over the last two decades, according to inverse modelling results (Tans et al., 1990; Ciais et al., 1995; Keeling et al., 1995; Denning et al., 1996a,b; Randerson et al., 1999); and isotopic analyses (Fung et al., 1997). While disagreements remain about the longitudinal distribution (Rayner and Law, 1999) and the processes responsible for this uptake (Holland et al., 1997a; Field and Fung, 1999; Houghton and Hackler, 1999), these results indicate that changes in the terrestrial carbon balance could be a significant factor in determining future trajectories of atmospheric CO2 concentration and thus the rate and extent of global warming; see further discussion in Chapter 3. The third generation LSPs have incorporated advances in our understanding of how green plants function, how they alter isotopic fractions of gases they come into contact with, and how they interact with radiation to produce distinctive signatures that can be observed by remote sensing satellites (Tucker et al., 1986; Sellers et al., 1992a; Myneni et al., 1995). As a result, LSPs can now be used to calculate mutually consistent land-atmosphere fluxes of energy, heat, water and carbon (Denning et al., 1996a,b; Randall et al., 1996).
Photosynthesis and stomatal conductance also exhibit strong diurnal variation, and improved representation of this has led to better simulation of the diurnal variation of surface heat and water fluxes and hence more realistic forcings for boundary-layer dynamics and convection (Denning et al., 1996a; Randall et al., 1996). These physiologically driven variations in the surface fluxes have a direct influence on the diurnal surface air temperature range in continental interiors and are directly sensitive to changes in atmospheric CO2 concentration (Collatz et al., 2000). Furthermore, increasing atmospheric CO2 is likely to have a direct effect on vegetation stomatal function through feedbacks in the photosynthesis-conductance system. Increased CO2 concentrations allow vegetation to maintain the same photosynthetic rate with a lower evapotranspiration rate. In a recent GCM study, tropical photosynthesis and transpiration rates were calculated to change only slightly under a CO2 concentration of 700 ppm, while the additional surface net radiation due to global warming was mainly returned to the atmosphere as sensible heat flux, boosting warming over the tropical continents by 0.4 to 0.9°C above the direct greenhouse warming of 1.7°C (Sellers et al., 1996a). It has been hypothesised that this effect may be partially countered by increased vegetation growth (Betts et al., 1997), but it is not clear to what extent this would be significant in already densely vegetated areas such as the tropical forests. To what extent can we trust these new models and their predictions?
Continues on next page
Other reports in this collection